NON-DESTRUCTIVE EVALUATION OF HORIZONTAL CRACK DETECTION IN BEAMS USING TRANSVERSE IMPACT

• Published in: Journal of sound and vibration Vol. 252; no. 2; pp. 343 - 360
• Main Authors: ISHAK, S.I., LIU, G.R., SHANG, H.M., LIM, S.P.
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 25-04-2002; Elsevier
• Subjects: Exact sciences and technology; Fundamental areas of phenomenology (including applications); Measurement and testing methods; Measurement methods and techniques in continuum mechanics of solids; Physics; Solid mechanics; Structural and continuum mechanics; Optical interferometry; Stratified material; Crack depth; Numerical method; Neural network; Modeling; Vibrometer; Inverse problem; Defect characterization; Laser interferometer; Optical measurement; Time domain method; Multilayer perceptrons; System identification; Localization; Non destructive test; Impact test; Crack
• ISSN: 0022-460X; 1095-8568
• DOI: 10.1006/jsvi.2001.4043

Numerical and experimental studies for crack detection in beam employing transverse impact are presented. In the numerical study, a beam model of wave propagation is developed to calculate the time history of beam response before, over and after the crack region. It is expected that the resulting wave in the beam will be scattered by the crack and will carry information on the location and geometry of the crack. Experiments using a scanning laser vibrometer on specimens containing simulated crack are then conducted to verify the numerical results. Comparison study between the numerical results and experimental observations are conducted; good correlation between theory and experiment is observed. The beam model of wave propagation and adaptive multilayer perceptron networks (MLP) are then used for inverse identification of crack parameters (i.e., crack location, depth and length) in the beams. Time-domain displacement responses calculated using the present beam model containing predetermined crack parameters are used as training data for the MLP. Once the MLP is trained, the MLP networks are then employed for inverse determination of an unknown crack in a beam using experimental displacement responses measured with a scanning laser vibrometer. Examples show that the procedure performs well for the determination of a wide range of values for the location, depth and length of the crack.